Tuesday, January 1, 2013
System Design: Mysterious Noises Re-Visited
Many years ago, I put together an index of airframe problems and published it as an article titled “Mysterious Noises.” A lot of people have used that information to good effect since then, but with so many technological changes, I thought it was over-due to re-examine the whole problem of how systems interfere with each other, and how this set of problems manifests in the airframe, especially as audio interference.
There are several paths for interference to occur between systems, it can happen by ground or power interaction, emission and susceptibility, transient behavior, magnetic interference, cable coupling, bad airframe positioning, poor equipment design and improper antenna loading. These situations combine and magnify in the small environment of any airframe to have some impact on all systems, and the balancing act of every avionics designer and installer is to minimize the effects so that airframe operation is as good as possible. Perfect operation under all circumstances is not a realistic outcome, as the implicit limits of the airframe generally prevent that, but it is possible to have system operation at a level that makes the shortcomings invisible and irrelevant to the flight crew.
To really understand the airframe environment and avionics systems, we have to start with an understanding of the primary power architecture, distribution and grounding structure. Power has to come from somewhere, and in an airframe, it is derived mechanically from the engine via generators or alternators, and distributed after regulation either via DC and AC busses and into a storage battery of some kind. Right at this moment, certain problems are injected into the airframe environment that have to be considered both as design issues and installation problems.
To understand how noise creeps into the audio system, let’s look at the basic power interconnect of the airframe, as this is where many problems first appear. Ideally, this is how we picture the power distribution takes place in a typical airframe:
Unfortunately, the real life system has many parasitic elements, and interactive systems. As composites displace aluminum, or distances increase, the ground return path becomes much higher in resistance, and more importantly, as ground resistance rises, each ground resistance in turn becomes a voltage created by changing load currents.
It is very common to see both AC and DC voltages impressed on the ground return, and values can be from millivolts to volts across the length of the airframe. Why is this important? If our audio signal is single ended, and shares a ground return path with these noises, then these signals mix with our audio signals and appear as noises, hums, buzzes, clicks and transients in the headset. Since a full microphone signal is only 150-250mV, it’s easy to see how even a few mV of noise can seriously contaminate the intercom and every outgoing transmission. In addition, every audio signal running through the ground also creates a related current and voltage, so this in turn causes cross-talk or the reception of unwanted audio signals when you thought they should be switched off.
The key to controlling noise in an aircraft audio system is to fully understand the cause of each type of noise, and to learn to identify them correctly. It is also very important to understand the limits of the system you are installing, and the impact of each wiring and routing decision you make during installation. The final results are the sum of all these factors, some within your control, and some always out of your control because of design limits, customer decisions and fundamental airframe limits. Very high performance is possible, but not always with every system, and not always with every ship.
These problems all fall into these two important groupings:
— Issues you can influence: system selection (if you pick it); wiring and routing
— Issues you can’t really influence: system selection (if the customer mandates it); composites; immovable locations and ambient noise
For the most part, audio system noise is literally anything you didn’t want or expect to hear. It can be in-band acoustically coupled noise from the ambient cockpit, or 400Hz AC accidentally coupled via the ships wiring. It can also be out-of-band AM RF transmissions detected and processed as demodulated audio, or cross-talk from an un-selected radio you thought was turned off. Most unexpected and objectionable to the flight crew are usually noises linked to other aircraft systems, such as alternators/generators, strobes, pumps and other electromechanical devices, which make their way into the audio stream via ground loop problems, and just never seem to go away, becoming perpetual audio noise thorns in your headset in every flight.
The shared ground signal path with ground loop noise voltages is the entryway for many audio problems, and it highlights why it is important to use a floating or balanced system for audio signals, rather than a single ended or grounded one to create the cleanest possible audio. This is not critical when only one audio panel or station is used, and if few radios or headsets are involved, but as multiple stations, users, and scattered radios appear, the ground loop noise begins to escalate quickly, and soon makes the system’s resulting high noise floor very irritating for the crew.
Audio system design plays a key role at this stage, because the ultimate ability to reject unwanted signals is firmly set at the design stage, not defined at the installation. In particular, factors like power supply noise rejection, cross-talk isolation and signal grounding set the base line performance, it can only deteriorate from there through bad installation practice. How significant can these design differences be? Single-ended designs typically have cross-talk and noise rejection figures of about 30dB, while floating systems typically achieve 60dB, or are a thousand times better at reducing noise and unwanted signals. This difference is not always serious with a single station, but quickly escalates when 2 to 4 stations are distributed through the airframe.
If your system is floating (the audio lines do not actually share a common path with the airframe ground), you can expect to see very significant reductions in ground loop and common mode noise, and if the design supports it, much lower cross-talk, especially in multi-station installs. Wiring technique comes into serious consideration at this point, as the final cross-talk and noise are now ultimately influenced by shielding and routing, no matter whether it is an analog, digital or a hybrid design.
Microphone lines are always the weakest part of any aircraft installation, because the signal level is so low, typically 150-250mVrms with a “carbon equivalent” microphone. Even very small amounts of signal contamination (literally 1-2mV) are quite audible, and will affect all radio transmissions and intercom operations. For this reason, correct grounding and shielding of these runs are of critical importance for over-all system performance. These lines are easily influenced both electrostatically and electromagnetically, which makes them very vulnerable to proximity cabling problems, especially AM RF lines, high current lines and 400Hz wiring. Low impedance dynamic (8-50 ohm) microphones are even more difficult to wire, as they have only ~250-700 microvolts of signal, and are very easily contaminated by stray magnetic coupling and RF. Use of special co-netic mu-metal foil shielded cable is usually required on these lines.
All systems start with an un-avoidable “noise floor,” which is the self-noise of the audio system itself, and which generally rises with increasing volume settings. You can only influence this factor by equipment selection. Next is the problem of acoustic ambient noise. Again you cannot usually alter the airframe, so your choices again involve the equipment selection if it has noise reduction properties, and particularly the headset/microphone selection. Considerable improvement in the user’s happiness is possible by making good choices at this stage.
Signals couple mainly electrostatically between wires (capacitive coupling), unless significant currents are involved, and shielded twisted pair wiring is very good at rejecting this adjacent signal coupling. Single conductor shielded wire is less effective because one of the conductors is also serving as the “shield.” Coupling primarily moves from high level signals to lower level ones, and increases in voltage as victim wire impedances go up.
If high currents are present, coupling is electromagnetic (speaker lines, inverter input currents or output currents, amplifier output or power lines, etc.). This inductive coupling is a serious problem, as conventionally shielded wiring is totally ineffective at stopping this type of interference. Special mu-metal co-netic shielded cable can provide some relief, but distance is the best solution to reduce coupling in these instances. High current lines should never be run with any low level audio wiring. Interestingly, the twisted pair shielded wiring does off both some capacitive and inductive coupling rejection, which makes it the most robust possible signal path. In fact, if you can do nothing else, and have no other tools or methods available, twisting audio line pairs together can often reduce coupling significantly.
No matter what topology you have, it can be seriously compromised by bad routing and wiring decisions, as once unwanted signals (in-band) are physically present in the audio system mixed with audio signals, they cannot generally be processed out. The most serous wiring/routing mistakes to avoid are these:
1. Never combine ANY 400Hz power or indicator wiring with ANY audio lines, even if shielded.
2. Never route AM Comm coax cables with low level audio wiring for any distance. This interference effect is called audio rectification, and occurs when AM RF signals encounter any non-linear junction (diode, transistor, etc.) and detect the envelope modulation, usually in a distorted way. If it is physically impossible to avoid tight cable coupling, use tri-axial cable (TRF-58 for example) for the RF feed line, with the outer shield grounded at one end only, this can dramatically reduce the coupling and interference effect.
3. Never route any high current wiring with audio lines. These high currents will magnetically couple, and conventional electrostatic shielding will not work, only distance is effective (remember the inverse square law).
4. Never run any audio line unshielded, unless the distance is short, and coupling is unlikely.
5. Do not ground mic/headset jacks to the airframe unless absolutely essential for some reason, as this will only increase ground loop noise. Return the signal commons to the audio system, and float the jacks.
One especially ugly combination of problems can occur with largely composite ships: the direct contamination of dynamic microphones (amplified or otherwise) to unblocked ambient RF. It is very difficult to shield microphones, and overhead RF sources with too much window or composite area surrounding the antenna sites can lead to direct RF injection into the microphone, which is basically a loop antenna. Lack of good airframe grounding and the RF transparency of composites increases both ground loop noise, and the opportunity for un-shielded direct RF interference. This is very hard to correct in a post-manufacturing environment, so we can only hope newer airframes will take these issues into better account to yield better avionics installations down the road.
Other radio-based problems that can manifest as audio interference include:
1. AM comms talking to each other, even when set to different frequencies, producing false sidetone or distorted audio. This can be complex to solve, but usually involves antenna re-location to opposite sides of the airframe, or to locations with no line-of-sight path.
2. FM comms coupling to each other harmonically, (150Mhz to 450MHz for example), producing phantom sidetone or distorted audio, or squelch breaking. Again, antenna relocation and breaking the line-of-sight may be required.
3. P-static discharge producing unwanted noise on ADF or AM comm radios. This is particularly troublesome from composite airframe structures, which are easily charged but not easily discharged quietly.
4. AM comms are highly susceptible to “rotor-modulation” caused by the rotating blades overhead in helicopters. This can make transmissions almost unintelligible, and garble incoming signals. Moving the AM Comm antenna to an underside surface with ground area between it and the blade system is usually the best solution.
The aircraft world is a strange mix of good and bad practices, with many hold-over techniques from 50+ years ago and early telephone technology. It also contains a large pool of older, marginal installations where everything was tied to the airframe, and where antenna siting was not well implemented. Sometimes, it is hard to dislodge these practices because of so much inertia and history.
To get the best system performance today, it is important to move past these early bad techniques, and look for areas where improvement is possible though better wiring techniques, better airframe layout and advanced equipment using floating audio techniques. The results can ultimately be very impressive.
Walter Shawlee 2 is the president of Sphere Research Corp. in West Kelowna, British Columbia, Canada, and a senior designer at Technisonic Industries. He can be reached at firstname.lastname@example.org.
To see a complete archive of Shawlee’s System Design columns, visit www.aviationtoday.com/shawlee.